Professional Documents
Culture Documents
IDC OPINION
IDC believes that software-defined storage (SDS) continues to transform IT deployment and
consumption of storage resources. For enterprises, SDS is really an on-ramp to deploying a hybrid
cloud — one that allows a metered on-demand consumption of private and public cloud resources.
Deploying SDS therefore is not a question of "if" but a question of "when." IDC's August 2015
Software-Defined Infrastructure Survey found that 75% of enterprises have deployed or are
considering deploying some form of SDS in their environment. For such enterprises, the question of
"why" is adequately clarified. In fact, IDC found that 61% of enterprises that deployed SDS have
realized tangible benefits such as reduction in capex/opex costs, ease of management, reduction in
provisioning time, and peace of mind knowing that they are no longer locked into a single vendor
solution. IDC believes that SDS is compelling enterprises to switch from a pattern of selecting systems
based on their capacity, performance, reliability, and cost characteristics to a service-focused
decoupled acquisition model in which hardware and software are acquired independent of each other.
Furthermore, the software in question (i.e., SDS) is not a one-size-fits-all proposition. In fact, for
enterprises and vendors to succeed with SDS, they have to take a use case–driven approach:
▪ Block-based SDS platforms serve as persistent storage for virtual machines, containers, and
physical servers used primarily for structured data sets and applications such as relational and
nonrelational databases.
▪ File-based SDS platforms, including distributed file systems that serve as repositories for
unstructured data, are used primarily for user and application data via standard datacenter-
based POSIX interfaces like NFS and SMB.
▪ Object-based SDS platforms are used for large globally dispersed repositories and data with
rich metadata requirements and are increasingly used for next-generation applications (NGAs)
that access data storage via RESTful APIs like S3 and Swift.
▪ Hyperconverged platforms are used in cases where compute and data layers need to be
adjacent — in use cases such as virtual server and virtual desktop infrastructure.
IBM deserves the spotlight for assembling a full-service SDS portfolio under the Spectrum family of
products. IBM's vision demonstrates that an incumbent storage vendor can be a leader in the nascent
SDS market. There is much work to be done by IBM and others to convince the rest of the market to
embrace SDS.
Enterprises are singularly focused on a digital transformation journey — one that requires a gradual but
consistent transformation in all aspects of how a firm operates. Maintaining a competitive edge is
mandatory and requires internal and external processes that allow the firm to bring new products and
services to market faster, provide an unparalleled customer service experience, and respond to market
trends in an agile manner.
Figures 1–4 (refer to the Appendix section) illustrate the findings from IDC's August 2015 Software-
Defined Infrastructure Survey. Findings include the following:
Storage is particularly critical because it is the only layer that deals with data persistence. Data is not
just handled as a transient entity; it "lives" here. Higher development costs associated with custom
hardware design and the resulting impact on the time to market of the solution have pushed suppliers
to shift their development efforts toward software solutions that run on industry-standard (aka
commodity off-the-shelf [COTS]) hardware. A common manifestation of this hardware in the datacenter
is in the form of x86-based servers with standard computing, networking and, of course, storage
components. In fact, one could argue that in the software-defined era, hardware platform commonality
translates to the delivery of computing, networking, and storage services via industry-standard servers.
Data persistence is a must for any solution to be classified as "storage," software defined or otherwise.
▪ Buyers have accepted that SDS forms the core of a building block strategy for next-generation
IT infrastructure. SDS allows IT to "future proof" infrastructure — as next-generation
applications come online, and the infrastructure can seamlessly adapt to the new workloads
and enable to-the-cloud initiatives. As companies move strategically to hybrid cloud, data
persistence will become more important.
▪ Buyers are aware that SDS is an approach to deliver at-scale storage using industry-standard
hardware. For most buyers, SDS offers two entry points for infrastructure acquisition: a
traditional capex/opex mode for traditional (corporate) IT initiatives and an agile but flexible
pay-as-you-go acquisition mode for line-of-business or project-driven IT initiatives.
▪ SDS is not and cannot be a one-size-fits-all approach. The choice of file-, block-, and object-
based SDS is governed by workload dependency. To that effect, buyers look for vendors that
can provide a full spectrum of SDS solutions that offer feature parity between standalone
installations and as a part of cloud frameworks like OpenStack.
Figure 5 (refer to the Appendix section) illustrates that the top 3 reasons for adopting SDS are a
tangible reduction in:
Figure 6 (refer to the Appendix section) illustrates that not all forms of SDS are used in the same
manner. File-based SDS is more popular, primarily due to the growth and proliferation of unstructured
user and analytics data. However, object-based SDS is catching on as enterprises adopt next-
generation applications. Block and hyperconverged trail behind, thus suggesting that buyers are not
yet there when it comes to embracing SDS for structured data sets.
IBM is approaching storage as a use case, not a product-by-product solution. A key attribute of IBM's
Spectrum Storage family is coherency. Unlike SDS-based offerings from other leading storage
suppliers, IBM's SDS portfolio (aka Spectrum Storage) is not a container of randomly acquired and/or
organically developed products. It is a well-thought-out framework with solutions that solve a set of
storage-related challenges in a consistent manner.
FUTURE OUTLOOK
Software-defined principles are driving the design of next-generation storage systems, as can be seen
from investments being made by incumbent and upcoming vendors. Increasingly, users are looking to
software-defined platforms as the medium to store data in a cost-effective manner, especially as data
sets get bigger. Enterprises have the opportunity to reduce deployment and operational costs by
aggressively rolling out a software-defined infrastructure paradigm, including SDS. The benefits
provided by SDS matter to these new workloads or NGAs. New deployment options are crucial to
expanding the use of SDS for NGAs. Accordingly, the datacenters of tomorrow will continue to change
with the proliferation of SDS platforms that:
▪ Offer flexible delivery models: From the compute layer to disk storage mechanisms and from
local open object interfaces to cloud-based interfaces, buyers need to have a wide range of
options for data storage. Initially, users will move their nonmission–critical and nonbusiness–
critical workloads to such platforms, and eventually, they will move more workloads to such
platforms.
CHALLENGES/OPPORTUNITIES
IDC's Software-Defined Infrastructure Survey asked respondents their reasons for not deploying SDS
or for getting rid of an installation. These responses illustrate that vendors still have some ways to go
in convincing their buyers to embrace SDS platforms. IBM's integrated solutions on the same SDS
framework may alleviate some of these concerns. Vendors, like IBM, must do more to educate their
customers on the benefits of SDS platforms in the areas of:
▪ Cost savings: Regardless of the manner in which the SDS platform is procured, it can deliver
substantial and measurable cost savings — both from initial acquisition savings and from
operational efficiencies.
▪ Ease of management: SDS platforms get a bad rap for being cumbersome to manage. Some
of it is perception, and some of it may be a function of bad implementation practices. Vendors
have to make their products easy to implement and easy to manage. More importantly,
vendors have to function as advertised, which means ensuring that there are no ambiguities
with respect to the selection of the platform for the workloads to be hosted on it.
▪ Vendor support: Some buyers believe that SDS platforms do not work as advertised. These
same buyers also think that SDS platforms do not receive the same level of support when
issues do occur. Vendors should do their best to debunk these misperceptions and provide
adequate assurances to such skeptical buyers.
▪ Ecosystem and application support: Buyers often hesitate to embrace SDS when there is a
lack of ecosystem and joint vendor certification from the application vendor. An ecosystem and
joint vendor certification does a lot to ensure a consistent message across the board.
SDS is here to stay. Both incumbent and up-and-coming storage suppliers are making a big push into
SDS. The value proposition of SDS platforms will become only more compelling as vendors focus on
this software-based delivery, away from custom hardware and flexible delivery models.
Buyers should continue to keep an open mind. By adopting newer software platform models that break
the traditional barriers between what are considered the compute, storage, and network components
of the infrastructure, they will be better positioned to support their business' digital transformation.
APPENDIX
FIGURE 1
0 5 10 15 20 25 30 35
(% of respondents)
n = 300
Note: Up to three business selections were permitted per respondent.
0 10 20 30 40 50
(% of respondents)
n = 300
Note: Up to three IT selections were permitted per respondent.
0 10 20 30 40 50 60
(% of respondents)
n = 300
Note: Multiple selections were permitted.
FIGURE 4
Unsure 2.7
0 10 20 30 40 50
(% of respondents)
n = 300
Note: Multiple selections were permitted.
Ease of management,
Ease of management
41.9
Reduced provisioning
Reduced provisioning times
times, 46.8
FIGURE 6
Block 24.2
File 31.5
Object 25.9
Hyperconverged 18.4
0 5 10 15 20 25 30 35
(% of respondents)
n = 302
Base = respondents who have currently deployed SDS platforms/systems
Global Headquarters
5 Speen Street
Framingham, MA 01701
USA
508.872.8200
Twitter: @IDC
idc-community.com
www.idc.com
Copyright Notice
External Publication of IDC Information and Data — Any IDC information that is to be used in advertising, press
releases, or promotional materials requires prior written approval from the appropriate IDC Vice President or
Country Manager. A draft of the proposed document should accompany any such request. IDC reserves the right
to deny approval of external usage for any reason.